COMPSCIX 415.2 Homework 3

library(tidyverse)
library(nycflights13)

3.8.1

• 1. This is an example of overplotting. So it under represents how many points are actually at each spot. I would use geom_jitter to show a better representation.

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) + 
  geom_jitter()

• 2. The paramters of geom_jitter that control the amount of jittering are width and height.

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) + 
  geom_jitter(width = 9, height = 0)

• 3. geom_count gives us different sized points depending on the amount of overlapping points. Both methods are ways to better view overlapping points. geom_count might have an area that has a large amount of points that can cover other points though, so it might not always be better to use it.

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) + 
  geom_count()

• 4. The default position for geom_boxpot is dodge. As you can see below, the plots look the same with and without position = "dodge".

ggplot(data = mpg) +
  geom_boxplot(mapping = aes(x = fl, y = cty), position = "dodge")

ggplot(data = mpg) +
  geom_boxplot(mapping = aes(x = fl, y = cty))

3.9.1

• 2. labs() allows you to modify the title, axis, legend, and plot labels. You can rename them.
• 4. The plot shows there is positive correlation between a car’s city mpg and highway mpg.
     As you can see without the coord_fixed() the x-axis distances are not the same as y-axis distances, so it can skew how people interpret data. Without coord_fixed() it looks like the city mpg increases faster than highway mpg.
     geom_abline adds in a line based on the parameters you enter into slope and intercept. As default, it makes the intercept = 0 and slope = 1. You can use it as a reference line when comparing the rest of the plot.

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) +
  geom_point() + 
  geom_abline() +
  coord_fixed()

ggplot(data = mpg, mapping = aes(x = cty, y = hwy)) +
  geom_point() + 
  geom_abline() 

4.4

• 1. The code doesn’t work because when you call it again, there is a typo. Instead of an “i”, there was a typo and they used “ı” instead.
• 2. Fixed Code

ggplot(data = mpg) + 
  geom_point(mapping = aes(x = displ, y = hwy))

filter(mpg, cyl == 8)

filter(diamonds, carat > 3)

5.2.4

• 1.1

filter(flights, arr_delay >= 120)

• 1.2

filter(flights, dest %in% c("IAH","HOU"))

• 1.3 We’ll first use airlines to get the list of airlines with their respective abbreviations.

airlines

filter(flights, carrier %in% c("UA","AA","DL"))

• 1.4

filter(flights, month %in% c(7,8,9))

• 1.5

filter(flights, dep_delay <= 0 & arr_delay > 120)

• 1.6

filter(flights, dep_delay >= 60 & arr_delay < dep_delay - 30)

• 1.7

filter(flights, between(dep_time, 0, 600))

• 2. between() gives us the values that are between the two arguments. It is inclusive. (I had already used it for 1.7. I guess I was supposed to use >= 0 & <= 600)
• 3. There are 8255 flights that have a missing dep_time. Other variables that are missing are dep_delay, arr_time, arr_delay, air_time, and occasionally tailnum. These might represent canceled flights.

filter(flights, is.na(dep_time))

• 4. NA ^ 0 is not missing because whenever you put anything to the zero power, it is always equal to 1. NA | TRUE is not missing either because the OR operator just requires one thing to be TRUE, and in this case, since one of the arguments is TRUE, it is not missing. FALSE & NA is not missing because for the AND operator, as soon as one argument is FALSE, the entire thing is FALSE.
     Unfortunately I cannot figure out the general rule due to NA * 0 giving us NA. Originally figured it was just if the expression always had the same result regardless of what you replace NA with, it would be fine, but anything multiplied by 0 should be equal to 0.

5.4.1

• 1. The most straightforward way would be to just list out the columns you want to select.

select(flights, dep_time, dep_delay, arr_time, arr_delay)

I had considered using contains() but doing this would end up grabbing sched_dep_time and/or sched_arr_time too. This same problem would arise from grabbing all the columns from dep_time to arr_delay.

We could also use the column numbers.

select(flights, c(4,6,7,9))

• 2. It returns the same result as if you only called it once.

select(flights, dep_time, dep_delay, dep_time, dep_time)

• 3. one_of() when used in conjunction with select, lets you select all the columns with names in a character vector. So in the example, you won’t have to type all the column names again, and you’ll get the year, month, day, dep_delay, and arr_delay columns.

vars <- c("year", "month", "day", "dep_delay", "arr_delay")
select(flights, one_of(vars))

• 4. I expected it not grab any columns due to none of the columns having the word “TIME” in all caps. By default it ignores case. In order to change the default you need to use ignore.case = FALSE

select(flights, contains("TIME"))

select(flights, contains("TIME", ignore.case = FALSE))